MECHANISMS BY WHICH CARDIAC
RESYNCHRONISATION THERAPY IMPROVES
CARDIAC PERFORMANCE IN HEART FAILURE
DR LYNNE KIRSTY WILLIAMS
A thesis submitted to
The University of Birmingham
for the degree of
DOCTOR OF PHILOSOPHY
Department
of
Cardiovascular Medicine
Clinical
and
Experimental
Medicine
The
University
of
Birmingham
University of Birmingham Research Archive
e-theses repository
This unpublished thesis/dissertation is copyright of the author and/or third parties. The intellectual property rights of the author or third parties in respect of this work are as defined by The Copyright Designs and Patents Act 1988 or as modified by any successor legislation.
Any use made of information contained in this thesis/dissertation must be in accordance with that legislation and must be properly acknowledged. Further distribution or reproduction in any format is prohibited without the permission of the copyright holder.
ABSTRACT
This thesis assesses the mechanisms by which biventricular and left ventricular pacing improves cardiac performance in patients with heart failure. We demonstrated for the first time that CRT results in an improvement in acute haemodynamic variables in heart failure patients with a narrow QRS duration that is comparable to the effects seen in heart failure patients with a broad QRS duration. In addition, we have shown that both biventricular (BIVP) and left ventricular pacing (LVP) significantly reduce external constraint to left ventricular filling, resulting in an increase in effective filling pressure. In heart failure
patients with evidence of external constraint at rest, the acute haemodynamic benefits of both BIVP and LVP were principally due to the relief of external constraint and preload
recruitment. However, in those patients with evidence of electrical dyssynchrony and a broad QRS duration, a significant haemodynamic benefit was derived from an enhancement in left ventricular contractility, presumably as a result of a reduction in left ventricular
dyssynchrony. Patients with external constraint appear to derive a greater haemodynamic benefit from pacing due to the significant increase in stroke work that is associated with relief of external constraint and preload recruitment, in addition to the increase in stroke work derived from enhanced contractility due to a reduction in dyssynchrony. These findings will inform better patient selection for this therapy and also optimisation of pacing strategy in individual patients.
ACKNOWLEDGEMENTS
Above everyone I would like to thank Dr Sue Ellery, my colleague and good friend, whose support and friendship have been unstinting for the past years. With her help and support I have navigated both the good and difficult times encountered during my period of research. As well as valuable and productive work, my time spent in research has resulted in this and many other good friendships.
I would like to thank my supervisor, Professor Michael Frenneaux, for his constant support, ideas and enthusiasm. He has played a major role not only in guiding me in my research, but also in providing invaluable advice and guidance for my future career in Cardiology. I would also like to thank Dr Kiran Patel, Dr Vince Paul, and Dr Paco Leyva for their invaluable role in these studies.
A great many thanks to all of the Cardiology secretaries, pacing technicians, and catheter lab staff in the departments were this work was carried out. Without their help, patience and guidance this work would never have been achieved.
Finally, I am indebted to both the British Heart Foundation and Medtronic Incorporated for the funding of my research, as well as for their continued support of research in Birmingham.
STATEMENT OF CONTRIBUTION
I undertook screening (with echocardiography, ECG and metabolic exercise testing) and recruitment of potential participants with a narrow QRS duration, as well as recruiting patients undergoing implantation of a biventricular pacemaker with a broad QRS duration.
I undertook the acquisition of the invasive haemodynamic data in 85% of all patients, and completed analysis of the data in all cases recruited. I was also involved in the collection and analysis of all echocardiographic data included in this thesis.
TABLE OF CONTENTS
CHAPTER 1: INTRODUCTION, LITERATURE REVIEW AND
HYPOTHESIS
INTRODUCTION
Heart failure burden and prognosis 1
The role of pacing in heart failure 2
The development of Cardiac Resynchronisation Therapy 3
Clinical studies of Cardiac Resynchronisation Therapy in heart failure 4 MECHANISMS OF BENEFIT FROM CARDIAC RESYNCHRONISATION THERAPY
Electrical and mechanical resynchronisation 7
Reduction in mitral regurgitation 9
Left ventricular reverse remodelling 11
Improvement in left ventricular filling 13
EXTERNAL CONSTRAINT AND DIASTOLIC VENTRICULAR INTERACTION 13
PATIENT SELECTION FOR CARDIAC RESYNCHRONISATION THERAPY 14
CURRENT CLINICAL GUIDELINES 15
SPECIFIC QUESTIONS TO BE ANSWERED 16
Narrow QRS duration heart failure 16
Biventricular versus left ventricular pacing 17
HYPOTHESES 18
CHAPTER 2: METHODOLOGY AND STUDY PROTOCOL
ASSESSMENT OF CARDIAC PERFORMANCE BY PRESSURE-VOLUME LOOPS 19
THE CONDUCTANCE CATHETER TECHNIQUE 20
GENERATED PRESSURE-VOLUME SIGNALS, LOOPS AND RELATIONS 23
ASSESSMENT OF SYSTOLIC FUNCTION 25
Indices of cardiac contractile function 25
End systolic pressure volume relation (ESPVR) 25
Preload recruitable stroke work relation (PRSWR) 26
ASSESSMENT OF DIASTOLIC FUNCTION 27
Indices of the active relaxation and left ventricular filling 27
End diastolic pressure volume relation (EDPVR) 27
ASSESSMENT OF EXTERNAL CONSTRAINT TO LV FILLING 28
ASSESSMENT OF LEFT VENTRICULAR SUCTION 30
Pressure volume loop analysis 30
Echocardiography 32
PATIENT RECRUITMENT 33
CHAPTER 3: ACUTE HAEMODYNAMIC EFFECTS OF CARDIAC
RESYNCHRONISATION THERAPY ON SYSTOLIC AND DIASTOLIC
FUNCTION
INTRODUCTION 37
METHODS 38
Patient group and baseline characteristics 38
Acute haemodynamic studies 40
Statistical analysis 41
RESULTS 42
Indices of left ventricular contractility and systolic function 42
Cardiac output 42
Left ventricular stroke work 43
Preload recruitable stroke work 44
dP/dtMAX 45
ESPVR (EES) 46
Indices of left ventricular diastolic function 47
dP/dtMIN 47
Tau 48
EDPVR (EED and KVED) 49
DISCUSSION 50
STUDY LIMITATIONS 53
CHAPTER 4: EFFECTS OF CARDIAC RESYNCHRONISATION THERAPY
ON EXTERNAL CONSTRAINT AND OPTIMAL INTERVENTRICULAR
TIMING DELAY
INTRODUCTION 55
METHODS 57
Patient group and baseline characteristics 57
Acute haemodynamic studies 58
Statistical analysis 59
RESULTS 60
Effect of pacing on external constraint 60
Diastolic ventricular interaction (DVI) 61
Left ventricular end-diastolic volume 62
Indices of systolic function in patients with and without DVI 63
Cardiac output 63
Left ventricular stroke work (LVSW) 65
Preload recruitable stroke work (PRSW) 67
dP/dtMAX 69
DISCUSSION 71
STUDY LIMITATIONS 75
CHAPTER 5: OPTIMAL INTERVENTRICULAR TIMING DELAY BASED
ON UNDERLYING PHYSIOLOGY AND QRS DURATION
INTRODUCTION 77
METHODS 77
Patient group and baseline characteristics 79
Acute haemodynamic studies 79
Statistical analysis 80
RESULTS 81
Effect of interventricular timing delay on pacing on dP/dtMAX 81
Effect of interventricular timing delay on LVSW 83
Optimal interventricular timing delay – external constraint 85 Optimal interventricular timing delay – underlying QRS duration 89
DISCUSSION 93
STUDY LIMITATIONS 96
CHAPTER 6: MECHANISMS BY WHICH CARDIAC
RESYNCHRONISATION THERAPY IMPROVES CARDIAC
PERFORMANCE IN HEART FAILURE
INTRODUCTION 98
METHODS 100
Patient group and baseline characteristics 100
Acute haemodynamic studies 101
Statistical analysis 105
RESULTS 106
Contribution to increased LVSW 106
Heart failure group as a whole 106
Narrow QRS group 108
Broad QRS group 110
Contribution to increased LVEDV 112
Heart failure group as a whole 112
Narrow QRS group 113
Broad QRS group 114
DISCUSSION 115
STUDY LIMITATIONS 117
CHAPTER 7: SUCTION FILLING OF THE LEFT VENTRICLE IN HEART
FAILURE AND THE EFFECT OF CARDIAC RESYNCHRONISATION
THERAPY
INTRODUCTION 120
METHODS 121
Patient group and baseline characteristics 121
Acute haemodynamic studies 121
Echocardiography 123
Statistical analysis 123
RESULTS 124
Pressure-volume loop versus echocardiographic parameters 124
Baseline suction filling parameters 126
Changes in suction filling with inferior vena cava occlusion 130
Changes in suction filling with CRT 131
DISCUSSION 133
STUDY LIMITATIONS 136
CONCLUSIONS 137
CHAPTER 8: DISCUSSION, CONCLUSIONS, AND FUTURE STUDIES
INTRODUCTION 138SUMMARY OF RESULTS 140
DISCUSSION 142
FUTURE STUDIES 146
CHAPTER 1
INTRODUCTION, LITERATURE REVIEW
AND HYPOTHESES
INTRODUCTION
The worldwide prevalence of heart failure is increasing in part due to an ageing population. In the developed world, heart failure affects 1-2% of the general population (1), causing about 5% of all adult hospital admissions, and complicating a further 10-15% (2). It is estimated that 0.2% of the population are admitted to hospital with this condition each year. In the developed world, the major aetiological factors are ischaemic heart disease, hypertensive heart disease, idiopathic dilated cardiomyopathy and valvular heart disease. Patients with advanced heart failure (New York Heart Association (NYHA) class III-IV), have both a poor quality of life and a poor prognosis, with one third or more of patients dying within 6 months of diagnosis. The annual mortality amongst those surviving beyond six months is 10-15% (2). Death for patients with advanced heart failure is usually due to disease progression, or less commonly is sudden.
There have been several therapeutic advances in the treatment of heart failure over recent years, with the use of beta-blockers, Angiotensin-converting enzyme (ACE) inhibitors, Angiotensin II-receptor blockers and mineralocorticoid receptor blockers. However the prognosis remains extremely poor and there is probably limited potential for therapeutic benefit from additional neurohumoral blockade.
For more than a decade several pacing modalities have been considered for patients with severe drug refractory heart failure, with varying results. More recently randomized clinical trials, observational studies and mechanistic studies have produced unequivocal support for the use of biventricular pacing in patients with NYHA III-IV refractory heart failure and evidence of ventricular conduction delay, most commonly seen as left bundle branch block on the electrocardiogram (ECG). These studies have consistently
demonstrated an improvement in functional status, quality of life and exercise capacity in these patients.
THE ROLE OF PACING IN HEART FAILURE AND THE DEVELOPMENT OF CARDIAC RESYNCHRONISATION THERAPY
Hochleitner et al (3) first described the use of conventional sequential AV pacing for the treatment of congestive heart failure. The authors initially reported that right-sided dual chamber pacing with a short programmed atrio-ventricular (AV) interval was of benefit in patients with heart failure who had no conventional indication for pacing, and
demonstrated a reduction in NYHA class, fall in cardiothoracic ratio on chest X-ray, and an increase in both systolic and diastolic blood pressure. The benefit was due to a reduction in pre-systolic mitral regurgitation, and hence an increase in left ventricular filling period. Pre-systolic mitral regurgitation is pronounced in those patients with prolongation of the AV interval and an elevated left ventricular end diastolic pressure. The acute
haemodynamic results in this and other studies (4;5) were impressive, however subsequent studies assessing effects on haemodynamic variables, symptoms and clinical outcomes in heart failure did not demonstrate significant benefits (6-8). More recently the ‘Dual
Chamber and VVI Implantable Defibrillator’ (DAVID) study (9) and the ‘Mode Selection’ (MOST) study(10) have suggested that right ventricular pacing per se may increase the
incidence of heart failure. It is believed that despite restoring AV synchrony, the attendant abnormal left ventricular electrical activation pattern results in dyssynchronous mechanical activity.
Over the past decade an emerging body of evidence has shown the utility of both biventricular (BIVP) and left ventricular (LVP) pacing. The concept of biventricular stimulation was first tested clinically by Bakker et al, who demonstrated that BIVP
improved functional capacity in patients with severe NYHA Class IV heart failure and left bundle branch block (LBBB) (11). The selection of patients with LBBB was based on the known adverse effects of LBBB, which results in late activation of the left ventricle (especially the left ventricular free wall) from the right ventricle via the septum, and subsequently a delay between the onset of left and right ventricular mechanical
contraction. The presence of LBBB also results in abnormal ventricular septal motion, which in turn is related to interventricular dyssynchrony and an abnormal pressure differential between the right and left ventricle. The abnormal septal motion results in an increase in the left ventricular end systolic diameter and reduction in the left ventricular ejection fraction, cardiac output, mean arterial pressure and dP/dtMAX. LBBB is associated with more severe symptoms of heart failure and also a higher mortality in heart failure patients (12;13).
In the mid-1990’s there was considerable controversy as to whether the restoration of AV synchrony, sequential RV stimulation, or biventricular pacing might play a role in the treatment of patients with heart failure and conduction disturbances. The PATH-CHF study was designed to prospectively address these questions. This multicentre trial tested the hypothesis that acutely optimised, chronic atrial–synchronous biventricular stimulation in
patients with Class III-IV heart failure and conduction disturbances would lessen
symptoms, increase exercise capacity, and improve quality of life. Results demonstrated a significant increase in functional capacity and quality of life, as well as a significant increase in anaerobic threshold and oxygen consumption at peak exercise (14). The initial clinical studies of long term biventricular pacing reported high peri-operative mortality figures, related to the early invasive epicardial lead placement procedure, involving thoracotomy under a general anaesthetic. Daubert et al, described the first transvenous approach for placing the left ventricular lead via the coronary sinus (15).
Several long term studies, of differing study design, have reported impressive results with biventricular pacing in heart failure patients with a prolonged QRS duration (>130ms) who are severely symptomatic (NYHA III-IV) despite optimal medication. The InSync study (16) reported significant improvements in exercise capacity, NYHA class and quality of life; the Multisite Stimulation in Cardiomyopathy (MUSTIC) study (17) reported a significant improvement in exercise tolerance, NYHA score, peak oxygen uptake and cardiac ejection fraction; the Multicentre InSync Randomized Clinical Evaluation study (MIRACLE) (18) reported a significant improvement in exercise capacity and quality of life, and a 50% reduction in hospitalisation in the first 6 months. Core centre analysis of the echocardiography data from the MIRACLE study also provides evidence of reverse remodeling of the left ventricle.
These studies were not powered to address mortality or morbidity as primary endpoints. However a meta-analysis of four studies (Contak CD, MUSTIC, MIRACLE, InSync ICD) with a total of 1634 patients reported a 51% reduction in progressive heart failure mortality from 3.5% to 1.7%, and a 21% reduction in hospitalization (19). The Comparison of
Medical Therapy, Pacing and Defibrillation in Heart Failure (COMPANION) study was a comparison of no pacing, biventricular pacing and biventricular pacing with ICD. It was halted prematurely because the ICD limb showed a 40% reduction in the combined end-point of death from or hospitalisation for heart failure, compared with a 34% reduction in the biventricular limb (20). This resulted in biventricular pacing being granted a class IIA indication by ACC/AHA/NASPE.
The results of CARE-HF have since been published. The trial was designed to evaluate the long-term effects of cardiac resynchronisation on the mortality and morbidity of patients with heart failure due to left ventricular systolic dysfunction with a QRS duration >
120msec who were already receiving optimal medical therapy. A total of 813 patients were randomised to device therapy or control and followed up for a mean of 29.4 months. The trial reported a 37% reduction in combined all-cause mortality (death) or unplanned cardiovascular hospitalisation, and a 36% reduction in all-cause mortality (21).
TABLE 1.1. Major trials of Cardiac Resynchronisation Therapy in heart failure
MUSTIC
INSYNC
MIRACLE
COMPANION
CARE-HF
Number n=131 n=103 n=453 n=1520 n=813 Design Randomised Controlled Prospective Observational Double-Blind, Randomised Controlled Randomised Controlled Randomised Controlled Criteria NYHA III EF < 35% QRS > 150 NYHA III/IV EF < 35% QRS > 150 NYHA III/IV EF < 35% QRS > 130 NYHA III/IV EF < 35% QRS > 120 NYHA III/IV EF < 35% QRS > 120 Outcome ↑ QOL ↑ EF ↑ 6MWD ↓ NYHA Class ↑ Peak VO2 max ↑ QOL ↑ EF ↑ 6MWD ↓ NYHA Class ↑ Peak VO2 max ↑ QOL ↑ EF ↑ 6MWD ↓ NYHA Class ↑ Peak VO2 max 34% ↓ (CRT) 40% ↓ (CRT-ICD) in death or hospitalisation due to CHF 24% ↓ (CRT) 36% ↓ (CRT-ICD) in death 37% ↓ in death / unplanned hospitalisation for a CVS event 39% ↓ in unplanned hospitalisation for a CVS event [6]
MECHANISMS OF BENEFIT IN RESPONSE TO CARDIAC RESYNCHRONISATION THERAPY (CRT)
Several mechanisms are thought to be responsible for the benefit seen in response to CRT in patients with heart failure.
ELECTRICAL AND MECHANICAL RESYNCHRONISATION
Dilatation of the left ventricle and associated fibrosis frequently induces intracardiac conduction delays resulting in dyssynchronous left ventricular motion. This often manifests as LBBB on the surface ECG. Biventricular pacing is currently advocated for patients with NYHA III and IV Class symptoms despite optimal medical therapy who have prolonged QRS durations (>120 ms). The latter is based on the original rationale that biventricular pacing acts via an improvement in cardiac ‘electrical synchrony’, hence the frequently used term “resynchronisation therapy”. The phenomenon of dyssynchrony is a consequence of a progressive, global, or focal disorder of the myocardium, leading to heterogeneous propagation of cardiac electrical and mechanical activity. There are at least three components to dyssynchrony that may impair cardiac function by affecting the systolic and diastolic properties of the heart.
(a) Prolongation of the atrio-ventricular (AV) conduction time leading to AV dyssynchrony is common in patients with heart failure (22). The delay in the onset of ventricular systole following passive and active ventricular filling gives rise to pre-systolic mitral regurgitation (23). This results in a lower left ventricular preload, higher pulmonary capillary wedge pressure, and a decreased cardiac output.
(b) Interventricular dyssynchrony occurs when the left and right ventricles fail to contract simultaneously. This is usually as a result of right or left bundle branch block, and often leads to septal contraction which is out of phase with one or other ventricular free wall.
(c) Intraventricular dyssynchrony occurs when there is marked heterogeneity in the timing of mechanical events between different segments of the left ventricle. As a consequence the ventricle expends a great deal of energy changing its shape but not ejecting blood. This dyssynchrony commonly extends to the papillary muscles resulting in mitral regurgitation. Although partly due to electrical dyssynchrony, other factors appear to contribute to mechanical dyssynchrony (24;25).
There is increasing evidence that there is only a weak correlation between electrical (QRS duration) and mechanical dyssynchrony and the benefit of CRT therapy (17;18;25;26). A substantial proportion of patients with a prolonged QRS duration (up to a third) do not exhibit inter- or intraventricular dyssynchrony (26;27), which may at least partly explain why up to a third of heart failure patients with prolonged QRS duration derive no benefit. Furthermore,
dyssynchrony is common even in heart failure patients with narrow QRS complexes (28), a group currently excluded from biventricular pacing treatment guidelines.
There is as yet no gold standard technique for quantifying intraventricular mechanical dyssynchrony. The QRS width on surface ECG is a simple method often used as a
surrogate for mechanical dyssynchrony, but the sensitivity for predicting the benefit from CRT is low (29;30). MRI can also detect areas of dyssynchrony, but this technique is expensive and cannot be repeated for follow-up after device implantation.
Echocardiographic tools are the most useful for the assessment of dyssynchrony. Studies have shown that measuring time to peak tissue velocity (Ts) from myocardial velocity curves using Tissue Doppler Imaging (TDI) is useful for quantitative assessment of systolic dyssynchrony. In particular, the dyssynchrony index (Ts-SD), or standard
deviation of Ts in a 12-segment model, was a powerful predictor of LV reverse remodeling
after CRT therapy (31). Tissue Synchronisation Imaging (TSI) is a technological advancement in the assessment of systolic dyssynchrony by transforming the Ts into different colour-coding. It has the advantage of providing a visual aid for quick
identification or regional delay within the LV wall. Yu et al have shown a good correlation between Ts or Ts-SD derived from both TDI and TSI (31). However, recent evidence suggests that conventional measures of dyssynchrony have high inter-observer variability (32) and may also underestimate the magnitude of dyssynchrony because they ignore radial dyssynchrony. Current dyssynchrony analysis is based on echo-Doppler methods which are largely derived from longitudinal motion data. This choice of orientation is mainly based on practical grounds given the available echocardiographic windows for transducer positioning. However, due to epicardial and subendocardial fiber orientation, cardiac contraction is principally radial. Helm and colleagues have demonstrated in a failing heart that dyssynchrony assessed by longitudinal motion is much less sensitive (33), suggesting that longitudinal motion data may not provide the most accurate and comprehensive means of assessing left ventricular dyssynchrony.
REDUCTION IN MITRAL REGURGITATION
Three mechanisms are responsible for mitral regurgitation in congestive heart failure. (a) As mentioned earlier, prolongation of the AV interval generates pre-systolic mitral
regurgitation by generating a diastolic ventricular-atrial pressure gradient. This is because the end of atrial contraction occurs much earlier than the onset of the rise in intraventricular pressure. Because of the altered geometry within a dilated left ventricle, appropriate mitral or tricuspid closure is probably not complete until the start of ventricular contraction. This can lead to early systolic or late-diastolic
mitral regurgitation. Brecker et al (4) showed that pacing with short AV intervals can reduce or abolish diastolic mitral regurgitation.
(b) ‘Functional’ systolic mitral regurgitation in dilated cardiomyopathy is a result of ventricular dilatation and increased chamber sphericity. The resultant increase in distance between the papillary muscles to the enlarged mitral annulus, as well as to each other, restricts leaflet motion and increases the force needed for effective valve closure (34-36). The mitral valve closing force is determined by the systolic left ventricle-left atrium pressure difference (or transmitral pressure gradient). Increasing this pressure gradient can decrease the effective regurgitant orifice area (EROA). Studies have shown that CRT (by re-coordinating LV contraction) acutely increases the maximal rate of LV systolic pressure rise (dP/dtMAX ) and thus the transmitral pressure gradient (37;38). Breithardt et al (39) studied 24 consecutive patients with severely impaired LV function, LBBB, and functional mitral regurgitation. CRT acutely reduced the severity of mitral regurgitation by
decreasing the effective regurgitant orifice area (EROA). This effect was directly related to an improvement in LV systolic function causing an accelerated rise in the transmitral pressure gradient (TMPG). The acute effect was independent of
geometrical changes (reverse remodelling).
(c) Kanzaki et al (40) have described a further mechanism for the reduction in mitral regurgitation seen with CRT. Mitral regurgitation in patients with a LBBB relates to an altered systolic balance of forces acting on the papillary muscles due to uncoordinated regional LV contraction. Utilising strain imaging on
echocardiography, they demonstrated a delay in peak strain at the mid-lateral segment adjacent to the anterolateral papillary muscle (implying tethering of the mitral leaflet), which was improved immediately after CRT.
FIGURE 1.1. Mechanisms of Improvement in Mitral Regurgitation Restoration of AV synchrony with an improvement in LV filling Ventricular reverse remodelling and a reduction in the sphericity index Increased TMPG and a reduction in the EROA Restoration of the systolic balance of forces acting on the papillary muscles
REDUCTION IN MITRAL REGURGITATION CARDIAC RESYNCHRONISATION THERAPY
LEFT VENTRICULAR REVERSE REMODELLING
Biventricular pacing has been shown to have a beneficial effect on left ventricular remodelling. Enhanced ventricular ejection efficiency resulting from coordinate contraction, a reduction in mitral regurgitation, and a reduction in
sympathetic/parasympathetic imbalance may contribute to chronically unload both ventricles and initiate reverse remodelling.
(a) Effects on left ventricular chamber geometry and dimensions
In a non-randomised study, Yu et al evaluated 25 NYHA class III-IV patients with an ejection fraction <40% and a QRS duration >140ms, who received biventricular pacing. Continued pacing at 3 months in these patients resulted in significant improvements in ejection fraction, left ventricular volumes, dP/dtMAX, myocardial performance index and the degree of mitral regurgitation, as well as a decrease in the
sphericity index. Withholding pacing for a 4 week period resulted in a progressive, but not immediate, loss of effect. It was concluded that the improvement in left ventricular mechanical dyssynchrony was the mechanism for this (41). Data from several
randomised trials consistently shows that CRT is able to induce reverse remodeling in the vast majority of patients. Stellbrink et al analysed 25 patients included in the PATH-CHF trial, focusing on changes in LV volumes after CRT. Left ventricular end diastolic (LVEDD) and end-systolic dimensions (LVEDS) were significantly reduced after 6 months, as were LV volumes (42). These findings are supported by data from the MIRACLE, CONTAK-CD, Vigor-CHF and INSYNC-ICD trials.
(b) Effects on neurohumoral pathways
Sympathetic activity is increased in patients with congestive heart failure, as demonstrated by elevated levels of circulating noradrenaline and an increase in adrenergic nerve outflow, as measured with microneurography. The cause of this increase is thought to be in part due to abnormal baroreflex control of adrenergic flow from the central nervous system. Although initially compensatory, chronic adrenergic activation promotes disease progression. Saxon et al recently reported no
significant alteration in noradrenalin levels during three months of biventricular pacing (43). However, Hamdan and colleagues compared BV pacing with single-site pacing in patients with LV dysfunction. Results demonstrated that both acute BV and LV pacing reduced sympathetic activity compared with RV pacing regardless of the QRS duration (44), and Braun et al demonstrated significant decreases in noradrenalin and NT-proBNP after short-term CRT in patients with congestive heart failure and conduction delay, although this effect was attenuated after 24 months of long-term follow-up (45).
IMPROVEMENT IN LEFT VENTRICULAR FILLING
Cardiac resynchronisation therapy may improve the diastolic filling time of the left ventricle in several ways:
1. Reducing pre-systolic mitral regurgitation
2. Improving ventricular synchrony during active relaxation, biventricular pacing might theoretically improve diastolic filling, although two studies have reported no improvement in Tau (25;38)
3. Reducing external constraint to left ventricular filling
EXTERNAL CONSTRAINT AND DIASTOLIC VENTRICULAR INTERACTION In 1895, Otto Frank (46) observed that as the left ventricle was stretched its force of contraction increased. Subsequently, Starling (47) reported that the force of ventricular contraction increased with increasing atrial pressure (the Starling relation), and later proposed a descending limb of the Starling curve (i.e. indicating a reduction in myocardial contraction at very high filling pressures). In an editorial in 1965, Katz (48) concluded that a descending limb of the Starling curve did indeed exist in severe heart failure, and had serious prognostic consequences. Data that have emerged since that time suggest that this descending limb relates to DVI in heart failure. Dupuis et al (49) reported a sustained haemodynamic benefit in some patients with severe chronic heart failure from a 72 h infusion of nitroglycerine. Among patients whose stroke volume increased, a tendency towards an increased LVEDV was noted, despite a fall in pulmonary capillary wedge pressure. Among patients in whom the stroke volume fell, however, LVEDV generally dropped.
Atherton et al (50) and Dauterman et al (51) have shown that in many heart failure patients the filling of the left ventricle may be impeded by external constraint from the raised pressure in the
stretched pericardium (pericardial constraint) and from the right ventricle via the interventricular septum (diastolic ventricular interaction). In these circumstances reducing central blood volume may acutely increase left ventricular diastolic volume despite reducing LV diastolic pressure. This is because the reduction in external constraint from the right ventricle and pericardium increases the true left ventricular distending pressure. LV pacing induces a phase shift such that LV contraction and filling both occur before they do in the RV (52-54). Because pericardial stretch (and therefore pericardial pressure) depends on total cardiac volume, a smaller RV volume during LV filling would result in less constraint to LV filling, a greater LV end-diastolic volume, and (by the Frank Starling mechanism) a greater LV stroke work. Bleasdale and colleagues (55) recently showed that LV pacing reduced this external constraint, effectively recruiting preload, presumably by causing the LV to fill relatively earlier, and therefore reducing the external constraint acting on the left ventricle at mid-diastole.
PATIENT SELECTION FOR CARDIAC RESYNCHRONISATION THERAPY
The present guidelines for patient selection for CRT are based largely on the entry criteria for the aforementioned larger clinical trials. The present criteria, as published by the National Institute for Clinical Excellence (NICE, 2003) are: drug refractory symptoms, NYHA III-IV, ejection fraction <35%, QRS duration >120ms, non-reversible cause, left ventricular end diastolic dimension >6cm. Despite the positive results of CRT trials, more than a third of patients receiving CRT therapy are non-responders. As mentioned earlier, echocardiographic markers of mechanical dyssynchrony show considerable variation in patients with broad QRS duration, and marked mechanical dyssynchrony may be present in heart failure patients with narrow QRS duration (25;28). Biventricular pacing (which reduces QRS duration) reduces indices of mechanical dyssynchrony, but curiously, despite increasing QRS duration, LV pacing also reduces mechanical dyssynchrony (25).
Furthermore, the BELIEVE study reported comparable clinical benefit from left
ventricular and biventricular pacing (56). Nevertheless, several studies, using a variety of techniques have shown that biventricular pacing produces the greatest echocardiographic, haemodynamic and clinical benefit in those patients with the greatest mechanical
dyssynchrony at baseline and reductions in dyssynchrony with pacing also predict these benefits (41;57;58). Some patients with narrow QRS complexes (who do not fulfill current criteria) may benefit. A better understanding of the mechanisms of benefit of biventricular and left ventricular pacing may lead to improved patient selection and optimisation of pacing strategy in individual patients.
TABLE 1.2. Current Guidelines for CRT Therapy
NICE GUIDELINES •Drug-refractory, NYHA Class III-IV •LVEF < 35%
•Sinus rhythm
•QRS Duration >150ms OR QRS >120ms and mechanical dyssynchrony on echocardiography •LVEDD > 6cm
•Non-reversible cause
ACC/AHA/NASPE GUIDELINES –
CLASS IIA INDICATION
•Drug-refractory, NYHA Class III-IV •Dilated or ischaemic cardiomyopathy •QRS duration >130ms
•LVEDD ≥ 55mm •LVEF ≤ 35%
SPECIFIC QUESTIONS TO BE ANSWERED
Aside from the haemodynamic benefits from cardiac resynchronisation therapy, many questions still remain:
(a) CRT in Patients with Heart Failure who have a Narrow QRS Duration
Previous small studies in patients with a narrow QRS duration have focused on clinical and echocardiographic end-points. Although these have shown a benefit from CRT based on clinical and echocardiographic parameters, these studies included QRS durations of < 150ms as ‘narrow’ and were limited to patients with evidence of dyssynchrony on echocardiography. In a study by Achilli et al (59), the authors report the efficacy of CRT in such a group of patients, with 14 of the 56 patients having a QRS duration of < 120ms. These findings were supported by those of Gasparini and colleagues (60), who reported an improvement in clinical and echocardiographic parameters in patients with a narrow QRS (13 of whom had a QRS < 120ms). The studies by Bleeker et al (61) and Yu et al (62) included 33 and 51 patients respectively with a QRS < 120ms. They showed an improvement in clinical parameters and
evidence of left ventricular reverse remodeling, but in addition the study by Yu et al showed an improvement in exercise capacity as evidenced by an improvement in maximal metabolic equivalent achieved on treadmill testing (62).
However, a recent much larger study (RETHINQ) in an ICD population who all met conventional dyssynchrony criteria for CRT but had QRS duration <130msec did not report a significant improvement in exercise capacity or evidence of reverse
remodeling (63).
(b) Biventricular (BIVP) Versus Left Ventricular Pacing (LVP)
Despite marked differences in ventricular pacing-induced QRS widths, BIVP and LVP pacing have proved equally effective in enhancing cardiac function in failing human hearts with LV dyssynchrony (37;38;64;65). This similar haemodynamic improvement despite striking differences in QRS duration supports the concept that mechanical rather than electrical resynchronisation is likely to be the most important phenomenon in achieving benefit from pacing therapy.
In order to clarify the respective mechanisms of BIVP and LVP responsible for similar haemodynamic improvement despite differing electrical activation, Bordachar et al (38;66) conducted an acute echocardiographic study in 33 severe heart failure patients to compare haemodynamic and dyssynchrony parameters during the two modes of pacing. Results showed that although LVP and BIVP produced similar haemodynamic improvements, LVP was associated with a substantial reduction of intra-LV
dyssynchrony. However, it was also associated with a shortened LV filling time and a longer aortic pre-ejection delay (time to onset of aortic flow on continuous wave Doppler), as well as worsened interventricular dyssynchrony. However, to date only intraventricular and not interventricular dyssynchrony has been associated with an adverse outcome in patients with heart failure (38;67;68).
As mentioned earlier, one of the mechanisms of improvement with CRT in heart failure is a reduction in external constraint. LVP induces a phase shift such that LV contraction and filling both occur before they do in the RV. Whether BIVP may similarly reduce external constraint is uncertain. Because it would be expected to produce a lesser phase shift in the time of ventricular filling, the benefit may not be as great.
HYPOTHESES
(1) The acute haemodynamic benefits of LVP and BIVP are similar in patients with heart failure and narrow QRS complexes to those seen in patients with broad QRS complexes.
(2) The acute haemodynamic benefits of LVP are principally due to a reduction in diastolic ventricular interaction.
(3) BIVP reduces diastolic ventricular interaction (in those in whom it is present) in patients with broad QRS complexes but not in those with narrow QRS complexes.
AIMS OF THE STUDY
Despite the wealth of data available on the effects of Cardiac Resynchronisation Therapy in congestive heart failure, a number of important questions remain. This study will assess the mechanisms by which biventricular and left ventricular pacing improves cardiac performance in patients with heart failure. It will assess the relative contributions of reduced diastolic ventricular interaction, enhanced contractility, and a reduction in mitral regurgitation to the acute benefit of each pacing modality in heart failure patients with both narrow and broad QRS complexes. The findings will better inform patient selection and also allow optimisation of pacing strategy in individual patients according to underlying physiology.
CHAPTER 2
METHODOLOGY AND STUDY PROTOCOL
ASSESSMENT OF CARDIAC PERFORMANCE BY PRESSURE-VOLUME LOOPS
Assessment of left ventricular systolic and diastolic pump properties is fundamental to advancing our understanding of cardiac pathophysiology and assessing the effects of both pharmacologic and non-pharmacologic therapies, particularly in heart failure. The use of the ventricular pressure-volume relation began as far back as 1895, when Otto Frank first described the relationship between filling volume and strength of myocardial contraction (46). However, the technique proceeded at a slow pace thereafter due to difficulties with assessing ventricular volumes in the intact human or animal heart(69;70).
With the development of the isolated blood-perfused canine heart model there was a resurgence of activity in the 1970’s and 1980’s (71;72). This led to a series of pivotal studies detailing the use of end-systolic and end-diastolic pressure-volume relations (ESPVR and EDPVR) as a measure of intrinsic myocardial pump properties (72;73). Further studies have described the characteristics of these relationships in detail, and have validated techniques for measuring ventricular volume (74), allowing these techniques to be applied to basic and clinical research. In addition, the physiological significance of the ESPVR and EDPVR has been demonstrated by their correlation with myocardial energy demand (75;76). Because the technique is applicable to the hearts of all species, pressure-volume analysis has become a standard technique in studies of both humans and animals of all sizes.
When assessing cardiac performance, it is important to remember that two distinct properties of cardiac function, that are intimately interrelated, are being assessed. One is assessment of the properties of the ventricle as a haemodynamic pump (both systolic and diastolic), and the second intrinsic properties of the myocardium itself. However, pump function is dependent on intrinsic myocardial properties such as muscle mass, myocardial architecture, and chamber geometry. The pressure-volume relation is used mainly to assess ventricular pump function, but some parameters derived from this relation yield
information about intrinsic myocardial properties.
THE CONDUCTANCE CATHETER TECHNIQUE
The conductance catheter technique developed by Baan et al (74;77) enables continuous real-time measurements of left ventricular (LV) volume and pressure. The method has been described extensively as a method of assessment of global systolic and diastolic ventricular function. The conductance methodology is based on the measurement of electrical conductance of the blood within the left ventricle. The catheter contains multiple electrodes which when positioned along the long axis of the left ventricle generate an intra-cavitary electric field (utilising the two most distal and two most proximal electrodes) and sense the resulting voltage gradients. The remaining electrodes measure segmental
conductance signals which represent the instantaneous volumes or corresponding slices of the left ventricle (Figure 2.1).
FIGURE 2.1. Positioning of conductance catheter in the left ventricle
The measured conductance (applied current divided by the measured voltage gradient) is subsequently converted to an absolute volume by taking into account the specific
conductivity of the blood and the electrode spacing. Parallel conductance (the offset in volume due to structures surrounding the left ventricular cavity i.e. the right ventricle) is determined by the hypertonic saline dilution method (74;78), and subsequently subtracted.
The conductance-derived stroke volume and cardiac output are generally an underestimation of actual stroke volume and cardiac output due to electrical field inhomogeneity and because the segments do not fully cover the LV long axis.
This is corrected by using a slope factor (α) which is calculated by comparing the catheter-derived cardiac output, and the cardiac output determined using the Fick equation:
___________O2 consumption ____________
Fick cardiac output =
((arterial sats) – (mixed venous sats)) x Hb x 1.34 x 10
Hence:
Slope factor (α) = Catheter derived cardiac output / Fick cardiac output
The catheter also contains a solid-state, high-fidelity pressure sensor to measure
instantaneous left ventricular pressure. Currently most pressure-volume studies performed in humans use combined pressure-conductance catheters. These catheters are typically 7-French, over-the-wire, pigtail catheters which are connected to a dedicated patients module to allow generation of an electric field, measurement of resting voltages, and the
acquisition of pressure, volume and ECG signals.
GENERATED PRESSURE-VOLUME SIGNALS, LOOPS AND RELATIONS
When positioned in the long-axis of the LV, the combined pressure-conductance catheter yields real-time segmental volume signals and an LV pressure signal with a temporal resolution of 4ms. The total LV volume is then calculated as the instantaneous sum of the segmental volumes. In order to assess ventricular pump function, pressure and volume signals are combined to construct a pressure volume loop, with each loop representing one cardiac cycle. The distinct cardiac phases are shown in Figure 2.2. Important parameters of both systolic and diastolic ventricular function can be determined either directly from the pressure-volume loop, or from the pressure and volume time curves and their derivatives.
FIGURE 2.2. A representative pressure-volume loop. Cardiac phases and timing of aortic and mitral valve opening and closure are depicted
A. Mitral valve closure; B. Aortic valve opening; B. C. Aortic valve closure; D. Mitral valve opening
Many of these parameters are to a large extent load-dependent, and in order to overcome this problem it is possible to construct pressure-volume relations which are much less load-dependent. This requires the acquisition of a series of pressure-volume loops over a range of loading conditions, which must be achieved by an intervention that has minimal effect on intrinsic myocardial function. This can be achieved utilising inferior vena caval balloon occlusion, which results in a rapid, purely mechanical, reduction in preload which is easily and rapidly reversible. A typical example is shown in Figure 2.3.
FIGURE 2.3. Pressure-volume loops acquired during preload reduction by means of inferior vena caval occlusion.
0
20
40
60
80
100
120
30
40
50
60
70
LV Volume ESPVR LV Pressure EDPVRESPVR – End-systolic pressure-volume relation EDPVR – End-diastolic pressure-volume relation
ASSESSMENT OF SYSTOLIC FUNCTION
INDICES OF CARDIAC CONTRACTILE FUNCTION
Important parameters characterising left ventricular contractile function can be determined directly from the pressure-volume loop without the need for construction of curves or relations between variables. These include stroke volume, cardiac output, stroke work, and dP/dtMAX. An important limitation of all of these indices is their relative load-dependence.
END-SYSTOLIC PRESSURE VOLUME RELATION (ESPVR)
As demonstrated by Figure 2.3, the ESPVR is constructed by connecting the pressure-volume points of each loop acquired during IVC occlusion. The ESPVR characterises the properties of the left ventricle at the point of maximal activation, and was initially thought to be a reasonably linear relation with a slope (EES – end-systolic elastance) and an
intercept on the volume axis (VO). Inotropic agents that increase contractile state have been shown to increase EES with relatively little change in VO (i.e. an increase in the slope of the ESPVR), while negative inotropes decrease the EES with little effect on VO (79). Similarly, increases in heart rate result in an increase in EES via the force-frequency relationship.
In the latter studies the end-systolic pressure volume relation (ESPVR) was determined using the single-beat method (EES) adopted from Takeuchi et al (80). There are several limitations to the assessment of ESPVR. Firstly, the ESPVR has been shown to be influenced by afterload impedance, but this effect is relatively small and can generally be ignored in most in-vivo studies as afterload conditions do not vary over very wide ranges. Secondly, the ESPVR is in reality non-linear over the entire range of pressures, and becomes convex in situations of enhanced inotropic state and concave with depressed inotropic state (81). Thirdly, ESPVR is dependent on muscle mass and geometry, but this
effect can be ignored in acute studies which assess an intervention and these variables are fixed over the various interventions.
PRELOAD RECRUITABLE STROKE WORK RELATION (PRSWR)
The PRSW relation is generated by plotting stroke work against end-diastolic volume during preload-reduction, resulting in a slope (Mw) and an axis intercept value (which is represented by the theoretical unstressed volume of the heart at a point when stroke work is equal to zero) (Figure 2.4). The theoretical unstressed volume of the heart is calculated using the equation LVEDV multiplied by a constant value k=0.72 for hearts with an end diastolic volume ≥ 95.7ml (82). This relation has many advantages over assessment of the ESPVR, including linearity, afterload independence, heart rate independence, and an independence of heart size, and presents the gold-standard for assessing changes in myocardial contractile performance.
FIGURE 2.4. Preload recruitable stroke work relation from a study patient during no pacing (OFF), biventricular pacing (BIVP) and left ventricular pacing (LVP)
ASSESSMENT OF DIASTOLIC FUNCTION
INDICES OF ACTIVE RELAXATION AND LEFT VENTRICULAR FILLING Important parameters characterising left ventricular relaxation and filling can be determined directly from the pressure data, although these too are limited by their dependence on load. These include dP/dtMIN (minimal rate of pressure change) and Tau (τ), the relaxation time constant.
END DIASTOLIC PRESSURE VOLUME RELATION (EDPVR)
The EDPVR is intrinsically non-linear due to different types of structural fibres being stretched in different pressure-volume ranges (83). In the low pressure-volume range compliant elastin fibres and myocytes with sarcomeric titin molecules account for diastolic stiffness (84), resulting in only a small increase in pressure for any given increase in volume. As volume is increased further, pressure rises more steeply as the slack lengths of collagen fibres and titin molecules are exceeded and further stretch is strongly resisted by these stiffer elements. Therefore chamber stiffness increases as end diastolic pressure or volume is increased. Nonlinear regression analysis may be applied to EDPVR data in order to develop simple indices of left ventricular chamber stiffness or compliance. Shifts of the EDPVR may reflect changes in myocardial properties (fibrosis, ischaemia or oedema) or pathological remodelling (with hypertrophy and chamber dilatation in heart failure), but take into account the net effects of changes in myocardial properties, chamber structural properties, and changes in the extracellular matrix.
The slope of the EDPVR can be determined using a single-beat method to calculate the EED (a measure of end diastolic chamber stiffness which is derived from a linear fit of the filling phase trajectory of the pressure-volume loop). Similarly the end-diastolic stiffness
constant KVED can be derived by fitting an exponential curve to the diastolic trajectory of the pressure-volume loop as previously described by Steendijk et al (85).
ASSESSMENT OF EXTERNAL CONSTRAINT TO LV FILLING
External constraint (EC) to LV filling can be determined using a modified static equilibrium technique wherebyexternal constraint is quantified as the difference in LVEDPbefore and after removal of the pericardium while a constantLVEDV is maintained. This technique can only be used when thechest is open. However, a modification can be applied by continuous measurement of LV pressure and volume duringocclusion of the IVC to acutely reduce RV volume and pressure (51).This acute reduction in RV volume removes external constraintto LV filling from the RV and
pericardium. During this intervention,LVEDP is progressively reduced over several beats. For eachbeat, the relation between LVEDP and LVEDV is assessed. Inthe absence of external constraint, IVC occlusion graduallyreduces both LVEDP and LVEDV, with the values progressing downwardand leftward along a single end-diastolic pressure-volume relation.In contrast, in severe heart failure, when marked external constraintis present, LVEDV initially (for a few beats) increases as LVEDPfalls. Only after the external constraint has been removed dothe pressure-volume values move down and to the left (Figure 2.5). The magnitude of external constraint is then calculated as shown in Figure 2.6.
FIGURE 2.5. Response to inferior vena caval occlusion in a patient with congestive cardiac failure and significant external constraint (left panel) and a healthy control (right panel)
FIGURE 2.6. Quantification of external constraint after inferior vena caval occlusion in a patient with congestive cardiac failure and significant external constraint
ASSESSMENT OF LEFT VENTRICULAR SUCTION
PRESSURE-VOLUME LOOP ANALYSIS
The term ‘suction filling’ has been defined by Katz as the phenomenon whereby the left ventricle relaxes faster than it can fill (86), resulting in a fall in pressure as the left
ventricle enlarges. Suction filling is therefore the period of filling which occurs during the period of pressure decline to minimum pressure after the mitral valve has opened. This can be calculated from the pressure-volume loop by the following equations:
Vsuction = V@Pmin – Vmin
% Suction filling = Vsuction____ Stroke volume
where Vsuction is the total volume of filling occurring between mitral valve opening (Vmin or minimum volume) and the point of minimum pressure (V@Pmin or volume at minimum pressure), and the percentage of suction filling is equal to Vsuction divided by stroke volume (see Figure 2.7).
FIGURE 2.7. Representative pressure-volume loop from a patient with congestive heart failure 0 10 20 30 40 50 60 70 80 150 170 190 210 230 250 270 290 Pre-systolic mitral regurgitation End-diastole Onset of suction End of suction End-systole Stroke volume Suction filling Non-suction filling IVRT [31]
ECHOCARDIOGRAPHY
Echocardiography was used to compare time intervals related to left ventricular filling with those obtained from the pressure volume loops. These recordings were obtained with the patients lying in a supine position at rest at identical heart rates to those during the pressure volume loop studies. Transmitral flow was recorded in the standard apical four-chamber view with the sample volume positioned between the tips of the mitral leaflets. The
isovolaemic relaxation time (IVRT) was derived by subtracting the time interval from Peak R-Wave of the ECG to mitral valve opening (onset of E wave) from the time interval from Peak R-wave to aortic valve closure. This method incorporating the use of the ECG complex eliminates the possibility of measurement error caused by valve artifact when measuring small time durations using continuous or pulsed wave Doppler imaging. The time interval from the Peak R-Wave of the ECG to the peak of the E wave on transmitral flow (T-Peak E) was also measured. The IVRT and T-Peak E time intervals derived from echocardiography was compared with the IVRT and T-Pmin derived from the pressure volume loops.
PATIENT RECRUITMENT
Patients were recruited from the heart failure clinics at the Queen Elizabeth Hospital (Birmingham), Good Hope Hospital (Sutton Coldfield), Sandwell Hospital (West Bromwich), and St Peter’s Hospital (Chertsey). Inclusion and exclusion criteria are detailed below:
Inclusion Criteria
Congestive heart failure due to ischaemic or dilated cardiomyopathy Sinus rhythm
NYHA III or IV breathlessness despite optimal tolerated medical therapy
Exclusion Criteria
Atrial fibrillation Structural valve disease Previous cardiac surgery Structural pericardial disease
Patient Groups
(1) Narrow QRS duration with no evidence of inter or intraventricular dyssynchrony (2) Broad QRS duration (unselected for dyssynchrony)
(3) Control group with structurally normal heart undergoing EPS and ablation
Details of patient recruitment are detailed in the following flow charts:
NARROW QRS PATIENT GROUP
Assessed for eligibility (n=94)
Excluded (n=52)
♦ Not meeting inclusion criteria (n=49) ♦ Declined to participate (n=2)
Failed Implantation (n=2) Proceeded to Implant of CRT device (n=42)
Unable to collect data at time of L + RHC (n=9) Underwent Left and Right Heart
Catheterisation for Haemodynamic Assessment (n=40)
Discontinued intervention (give reasons)
Analysed (n=31)
BROAD QRS PATIENT GROUP
Patients undergoing Implant of CRT device (n=30)
Unable to collect data at time of L + RHC (n=5) Underwent Left and Right Heart
Catheterisation for Haemodynamic Assessment (n=30)
Discontinued intervention (give reasons)
Data excluded from analysis due to poor quality signal (n=2)
Analysed (n=23)
STATISTICAL ANALYSIS
All data are expressed as the mean value ± SD. For continuous variables that were not normally distributed, the median and range are expressed.
A repeated measure ANOVA was used to assess changes for paired samples if the data was normally distributed based on a Kolmogorov-Smirnov test. For data that was not normally distributed, a Kruskal-Wallis test was used and medians reported. If the repeated measure ANOVA was statistically significant then pairwise comparisons were performed.
For comparison of data between patients groups (independent samples) a non-parametric Mann-Whitney test was performed.
Statistical significance was assumed at p<0.05.
CHAPTER 3
ACUTE HAEMODYNAMIC EFFECTS OF CARDIAC
RESYNCHRONISATION THERAPY ON SYSTOLIC
AND DIASTOLIC FUNCTION
INTRODUCTION
The effects of cardiac resynchronization therapy (CRT) in patients with heart failure and a QRS duration ≥ 120ms are well established (20;21).In the Cardiac Resynchronization Therapy Heart Failure (CARE-HF) study, CRT was associated with a 40% reduction in all-cause mortality (21). This and other studies have also shown that CRT leads to an
improvement in symptoms and a reduction in hospitalization. The predominant mechanism of benefit has been considered to be improvement in both inter and intra-ventricular (LV) dyssynchrony. However, additional mechanisms independent of resynchronisation contribute to the benefit derived from CRT.
Hemodynamic studies have previously demonstrated the acute (37;38;65;87;88) effects of CRT in patients with heart failure and a broad QRS duration.Data from studies by Kass et al (38) and Auricchio et al (37) demonstrated an acute 15% increase in LV dP/dtMAX in response to biventricular pacing. Steendijk et al performed pressure-volume loop studies at 6 months in patients who had undergone biventricular pacing, and demonstrated a 34% increase in LV stroke work and an 18% increase in LV dP/dtMAX in response to pacing (89). We previously showed that LV pacing produces an acute haemodynamic benefit (reduced pulmonary capillary wedge pressure, increased stroke volume) in patients with
heart failure and a QRS<120 ms (25), although these patients had not undergone any prior assessment for the presence or absence of dyssynchrony. Even patients with narrow QRS complexes who do not have dyssynchrony might be expected to benefit from LV and potentially biventricular pacing by relief of diastolic ventricular interaction if present.
In this study we assess the acute haemodynamic effects of both biventricular (BIVP) and left ventricular only pacing (LVP) in symptomatic heart failure patients with a QRS duration<120ms who did not meet the ‘conventional’ criteria for dyssynchrony and compared these effects with those demonstrated in symptomatic heart failure patients with a broad QRS duration who were unselected for the presence or absence of dyssynchrony.
METHODS
Patients. Thirty one patients with narrow QRS duration and 23 with broad QRS duration heart failure were recruited into the study. All patients had an LVEF≤35% as determined by echocardiography. Patients were in NYHA class III or IV despite optimal tolerated medical therapy that included diuretics and Angiotensin-converting enzyme (ACE) inhibitors. In addition, all patients with a narrow QRS ≤ 120ms had no evidence of inter- or intraventricular dyssynchrony. The former was defined as a Qp-Qa time delay > 40ms, whereas the latter was defined as a septal-posterior wall motion delay > 130ms, or an intraventricular septal-lateral wall delay > 40ms. Patients with a narrow QRS who met two or more of the above criteria were excluded. Baseline patient characteristics for the two groups are shown in Table 3.1.
TABLE 3.1. Baseline patient characteristics Narrow QRS (n=31) Broad QRS (n=23) Age 62 ± 14 years 69 ± 10 years Male 28/31 (90%) 19/23 (83%) Ischaemic cardiomyopathy 17/31 (55%) 14/23 (61%)
NYHA Class III
29/31 (94%) 21/23 (91%) NYHA Class IV 2/31 (6%) 2/23 (9%) LVEF 26 ± 5% 24 ± 7% Beta-blockers 25/31 (81%) 16/23 (70%) ACE-inhibitors / ARB 29/31(94%) 23/23 (100%) Spironolactone 25/31 (81%) 17/23 (74%) Diuretics 30/31 (97%) 23/23 (100%) Qp-Qa interval 15 ± 9ms N/A Septal-posterior wall delay
126 ± 54ms N/A Yu-dyssynchrony index 22 ± 7 N/A [39]
Acute Haemodynamic Studies. Acute haemodynamic studies were performed in the cardiac catheterization laboratory at the time of CRT device implantation with patients in the non-sedated and supine state. Catheterisation of the left ventricle was performed by a standard over-the-wire technique. The dual-field conductance catheter (CA-71103-PL catheter, CD Leycom, The Netherlands) was then positioned in the apex of the ventricle. We applied a modified parallel conductance calibration via a right atrial injection(78) to avoid catheterisation of the right ventricle or pulmonary artery.
All data were acquired during an unforced end-expiratory breath hold. From each
acquisition run, the derivatives of pressure and volume were calculated as the mean of the 10 to 15 consecutive beats free from atrial or ventricular ectopic activity. Pressure-volume analysis was also performed during an inferior vena caval (IVC) occlusion, which reduced central blood volume and RV pressure acutely, achieved with a 40-mm IVC occlusion balloon catheter (Meditec, Boston Scientific International). Data were acquired with a CFL-512 system (CD Leycom), which allows further offline analysis (CircLab, Leiden University, The Netherlands). The haemodynamic measurements were undertaken during no pacing (OFF), in biventricular (BIVP) and in left ventricular only pacing (LVP) modes with AV intervals set at 100msec, and interventions applied in a random order, each with a run-in (stabilization) period of 5 minutes.
The following parameters were derived from the pressure volume loops (PVL) at baseline (OFF), during inferior vena cava occlusion (IVCO), and in both BIVP and LVP pacing modes: dP/dtMAX and dP/dtMIN, absolute left ventricular stroke work (LVSW), cardiac output (CO) and left ventricular end-diastolic volume (LVEDV). A plot was constructed of beat-by-beat LV end-diastolic volume versus LVSW before and during IVC occlusion in
all pacing modes (the Preload Recruitable Stroke Work Relation or PRSWR). The end-systolic pressure volume relation (ESPVR) was determined using the single-beat method (EES), the end-diastolic pressure-volume relation (EDPVR) was determined using the single-beat method to calculate the EED and the end-diastolic stiffness constant KVED as described in Chapter 2.
In order to negate the effect of any change in heart rate on systolic function (via the force-frequency relation), a tracking VDD pacing mode was utilised to ensure a stable mean heart rate across all interventions (as shown in Table 3.2), with no statistically significant difference in heart rate between pacing modes.
TABLE 3.2. Mean heart rate (beats per minute) during different pacing modes
PACING MODE Narrow QRS (n=31) Broad QRS (n=23)
OFF 73 bpm 70 bpm
BIVP 76 bpm 70 bpm
LVP 76 bpm 72 bpm
Statistical Analysis. All data are expressed as the mean value ± SD. The Mann-Whitney rank-sum test was used to compare independent samples between the two groups. A repeated measure ANOVA was used to assess the effect of vena caval occlusion and pacing if the data was normally distributed based on a Kolmogorov-Smirnov test. For data that was not normally distributed, a Kruskal-Wallis test was used and medians reported. Statistical significance was assumed at p < 0.05.
RESULTS
Indices of left ventricular contractility and systolic function
Cardiac Output
Cardiac output was similar at baseline for the narrow and broad QRS groups (2.6 ± 0.8 vs. 2.7 ± 0.6 l/min respectively; p=0.92). Cardiac output increased by 25% in response to both BIVP and LVP in the narrow QRS group (p=0.02) and by 19% in the broad QRS group (p=0.04). There was no significant difference between BIVP and LVP in either group, or between the two groups.
FIGURE 3.1. The effect of biventricular (BIVP) and left ventricular pacing (LVP) on cardiac output (in l/min) compared with baseline values (OFF)
CARDIAC OUTPUT (litres/minute)
†
†
†
†
PACING MODE
Error bars indicate ± 1 SEM;
†
indicates a statistically significant increase from OFF (p<0.05)Left ventricular stroke work (LVSW)
LVSW was similar at baseline for the narrow and broad QRS groups (2292 ± 1060 vs. 2110 ± 629 ml*mmHg respectively; p=0.87). LVSW increased by 26% in response to BIVP and 31% in response to LVP in the narrow QRS group (p=0.04) and by 32% and 26% in the broad QRS group in response to BIVP and LVP respectively (p=0.03).
FIGURE 3.2. The effect of biventricular (BIVP) and left ventricular pacing (LVP) on LVSW (in ml*mmHg) compared with baseline values (OFF)
LVSW (ml*mmHg)
†
†
†
†
PACING MODE Error bars indicate ± 1 SEM;†
indicates a statistically significant increase from OFF (p<0.05)Preload recruitable stroke work (PRSW)
PRSW was similar at baseline for the narrow and broad QRS groups (38.6 ± 20 vs. 40.2 ± 13.4 respectively; p=0.45). PRSW increased by 34% in response to BIVP and 30% in response to LVP in the narrow QRS group (p=0.03) and by 37% and 29% in the broad QRS group in response to BIVP and LVP respectively (p=0.04). There was no significant difference between BIVP and LVP in either group, or between the two groups.
FIGURE 3.3. The effect of BIVP and LVP on PRSW compared with baseline (OFF) PRSW
†
†
†
†
PACING MODE
Error bars indicate ± 1 SEM;
†
indicates a statistically significant increase from OFF (p<0.05)dP/dtMAX
dP/dtMAX was significantly higher at baseline in the narrow compared with the broad QRS group (872 ± 189 vs. 690 ± 179 respectively; p<0.01). dP/dtMAX increased by 15% in response to BIVP and by 21% in response to LVP in the broad QRS group (p=0.03) and by 13% in the narrow QRS group in response to both BIVP and LVP (p=0.04).
FIGURE 3.4. Effect of BIVP and LVP on dP/dtMAX compared with baseline (OFF)
dP/dtMAX (mmHg/sec)
†
†
†
†
PACING MODEError bars indicate ± 1 SEM;
†
indicates a statistically significant increase from OFF (P<0.05)ESPVR (EES)
